iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
You can compose a Snowflake Cortex Agent with ServiceNow Knowledge Graph by exposing Knowledge Graph as a tool through ServiceNow’s MCP Server Console, then configuring Snowflake to connect to that remote MCP server. The platforms document these capabilities separately; the available documentation does not establish a vendor-certified, jointly validated Snowflake–ServiceNow deployment. Treat this as an integration pattern and confirm the endpoint, authentication, entitlement, and access policies in your own ServiceNow instance before implementation.
How the integration fits together
In this design, ServiceNow hosts the Knowledge Graph MCP tool and Snowflake Cortex Agent acts as the MCP client. Snowflake connects to the external server through an external MCP server object backed by an API integration. The agent can discover the exposed tools and call them when needed to obtain ServiceNow context.
Snowflake’s MCP connector documentation describes tool discovery using MCP tools/list and tool invocation using tools/call. ServiceNow’s Australia-release documentation, updated July 8, 2026, describes exposing Knowledge Graph functionality to MCP clients through the MCP Server Console. Those documented capabilities make the composition plausible, but do not by themselves confirm compatibility for a particular tenant, release, or OAuth configuration.
Free tools Windows power users keep installed
One-click scans. No signup required.
What to confirm in ServiceNow first
ServiceNow’s Knowledge Graph integration documentation says administrators can use Knowledge Graph as a tool in the MCP Server Console so agents can receive context and insights based on live instance data. It points to a separate “Create a tool from Knowledge Graph” procedure for adding and exposing the tool to MCP clients. Confirm that procedure and the resulting tool are available in the target instance; the integration overview does not specify every remote-client setup detail.
#1 Best Overall
ServiceNow’s Australia-release Knowledge Graph configuration page, updated March 12, 2026, describes Knowledge Graph as an AI Platform feature available with activation of a ServiceNow generative AI application. It says activation installs the Graph QL plugin and Knowledge Graph application by default, and that Knowledge Graph must be configured for Virtual Agent as part of adoption. These are configuration details, not confirmation that a specific instance is entitled to or configured for the MCP Server Console feature.
- Verify that the Knowledge Graph tool appears in the MCP Server Console and is exposed to the intended client.
- Obtain the exact MCP endpoint and supported authentication method from the instance administrator and current ServiceNow documentation.
- Confirm the OAuth client registration, authorization scope, user authorization requirements, and instance-side access policies. The Snowflake documentation does not specify ServiceNow’s endpoint format or OAuth scopes for this integration.
Configure Snowflake to connect to the remote MCP server
Snowflake’s documented external MCP connector flow has four stages. The examples cover other providers and custom MCP servers, not a first-party ServiceNow connector, so use the details verified for the target ServiceNow instance rather than assuming a standard ServiceNow configuration.
Rank #2
- Gather provider details. Obtain the remote MCP server URL and OAuth details from the ServiceNow administrator. Confirm the server’s authentication method and approved client configuration before creating Snowflake objects.
- Create the Snowflake API integration. Configure the API integration with the remote server URL, OAuth client credentials, and required endpoints using Snowflake’s current MCP connector instructions. Keep credentials under the organization’s approved secret-management and rotation practices.
- Create an external MCP server object. Configure the Snowflake object to reference the API integration. Confirm that both objects are enabled.
- Add the server to the Cortex Agent specification. Have the agent developer reference the external MCP server so the agent can discover and call its available tools.
- Authenticate users with the external service. Users authenticate through OAuth as required by the configured integration and the ServiceNow instance. Verify the user flow and authorization behavior in the target environment.
At runtime, Snowflake checks that both the external MCP server and its backing API integration are enabled before a tool call. The exact SQL syntax and OAuth values should come from Snowflake’s current connector documentation and the tenant-specific ServiceNow configuration; they are not established for this particular pairing here.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Set permissions and establish the trust boundary
Snowflake requires the connecting role to have USAGE on both the external MCP server and its backing API integration to connect and discover tools. Grant those privileges to the narrowest appropriate role. Separately, configure and test ServiceNow’s policies for the MCP tools and the records or Knowledge Graph data those tools can access; Snowflake object privileges do not replace ServiceNow authorization.
Snowflake states that external MCP servers are not provided, maintained, or verified by Snowflake. The organization therefore needs to assess whether the ServiceNow endpoint is trusted for the intended data, whether processing and disclosure are permitted, and which third-party terms apply. Assign operational owners for credentials, tool-call monitoring, access reviews, and changes to the endpoint or ServiceNow release.
Distinguish the remote ServiceNow server from Snowflake-managed MCP
The ServiceNow MCP server described here is remote from Snowflake and provides ServiceNow tools to the Cortex Agent. A Snowflake-managed MCP server serves a different direction: it can expose Snowflake capabilities to an MCP client. Snowflake’s managed-MCP guidance recommends exposing a Cortex Agent as the only client-facing tool when the goal is governed business questions. That is an optional, separate design choice, not a requirement for the agent to call ServiceNow.
| Design axis | Snowflake agent calling ServiceNow | Snowflake-managed MCP for another client |
|---|---|---|
| Client-facing orchestration | Cortex Agent | External MCP client |
| Tool host | Remote ServiceNow MCP server with Knowledge Graph tool | Snowflake-managed MCP server exposing Snowflake capabilities |
| Primary authorization boundary | Snowflake role privileges and OAuth integration, plus ServiceNow authentication and access policies | Snowflake-managed MCP configuration and the permissions governing the exposed Snowflake capabilities |
These are architectural axes, not mutually exclusive product alternatives: an organization could use each pattern for a different client or purpose.
Validate the integration before rollout
Run a controlled test with representative least-privilege identities before broad deployment. Check each layer independently so a discovery failure, authentication issue, or denied data access can be traced to the right system.
Quick Recap
Best Value
- In ServiceNow, confirm the Knowledge Graph MCP tool is available and exposed only to the intended client.
- Verify the exact endpoint, OAuth method, scopes, client registration, and user authorization behavior for the target instance.
- In Snowflake, confirm the API integration and external MCP server object contain approved settings and are enabled.
- Check that the Cortex Agent specification references the intended external MCP server.
- Verify the connecting role has only the required
USAGEprivileges on the server and API integration. - Test tool discovery and invocation with representative identities, and verify ServiceNow limits access to the intended tools and data.
- Observe tool calls and assign owners for monitoring, credentials, data-rights reviews, and changes to either platform’s configuration.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
